Collaborative Research: Living Building Information Model (BIM): A Layered Approach for Automatic and Continuous Built Environment Model Update

基础设施和建筑的设计具有很长的生命周期--大约几十年。世界上的许多建筑在经历了无数次的翻修努力后，在几个世纪后仍在运行。这一漫长的运营阶段代表了建筑生命周期的大部分时间，但有关维护和翻新的信息很少保持最新。建筑信息模型(BIM)可以通过集中存储这些数据来缓解这种数据短缺。然而，即使有了BIM，由于很难在建筑物的整个生命周期内持续手动更新，建筑物的更新也不会保持不变。这项研究将创造一种方法，通过利用机器视觉的最新进展来自动更新BIM。这一自动化过程可以系统地、持续地分析建筑环境，检测与先前评估相比的变化。它可以提取和更新关键建筑信息，并最大限度地减少人为错误和工作量。这项研究将导致建筑记录保存的根本变化，并使建筑运营商受益，使维护和翻新活动能够更容易地在建筑物的整个生命周期内进行规划。这项工作的结果将被整合到本科和研究生教育模块中。为YouTube制作的视频也将被开发，以吸引高中和代表性不足的人进入土木工程职业生涯。该项目将生成上下文-数据关系，供基于主动照明范围摄像机的机器视觉系统使用，该系统用于根据结构变化自动更新BIM数据库，并利用从BIM面向对象的数据库模型、计算机辅助设施管理(CAFM)数据库和人工操作员输入的专家知识中提取的元数据。该项目将解决几个智力方面的挑战。主要的挑战在于这项工作的机器视觉部分，它将需要在对象的3D几何模型之外提供额外的元数据，从而将机器视觉的边界推向土木工程系统的背景。另一个挑战是扩展构建环境的数据建模功能。具体地说，将创建从机器视觉转换元数据以识别和获得特定于对象的有意义的上下文数据的能力。还将开发一个具体的逻辑组成部分--背景决策制定者(CDM)，以合并来自多个数据源的元数据。最后，整个系统将在一个室内翻新项目中进行测试，以提供一个现实的机器学习过程，随着时间的推移，这个过程将变得更加健壮。

Infrastructure and buildings are designed to have long life cycles - on the order of decades. Many buildings in the world are still in operation after centuries amid numerous renovation efforts. This long operational phase represents the majority of a building's lifecycle, yet the information regarding maintenance and renovation is rarely kept up to date. Building Information Models (BIMs) can alleviate this data shortage by centrally storing this data. However, even with a BIM, building updates are not kept due to the difficulty of continuous manual updates over a building's lifetime. This research will create a method to automatically update a BIM by exploiting recent advancements in machine vision. This automation process can systematically and continuously analyze the built environment, detecting changes from a previous assessment. It can distill and update the critical building information with minimal human error and effort. This research will result in a fundamental change in construction record keeping and benefit building operators by enabling maintenance and renovation activities to more easily be planned throughout a building's lifetime. The findings of this work will be integrated into undergraduate and graduate educational modules. Videos created for YouTube will also be developed to attract high school and underrepresented persons to a career in civil engineering.The project will generate contextual-data relationships for use by an active illumination range camera-based, machine-vision system for automatically updating a BIM database with construction changes and leveraging metadata distilled from BIM object-oriented database models, the Computer-Aided Facilities Management (CAFM) database, and expert knowledge input by human operators. The project will tackle several intellectual challenges. A primary challenge resides in the machine vision component of the work, which will need to provide additional meta-data beyond a 3D geometric model of an object, pushing the boundary of machine vision for the context of civil engineering systems. Another challenge is to extend data modeling capabilities for the built environment. Specifically, capabilities to translate meta-data from machine vision to identify and obtain meaningful contextual data that is specific for the objects will be created. A specific logical component, the Contextual Decision Maker (CDM), will also be developed to merge meta-data from multiple data sources. Finally, the entire system will be tested in an indoor renovation project to provide a realistic machine-learning process that will grow more robust over time.